AI-enhanced design of carbon capture and utilization networks using surrogate-based superstructure optimization

초록

Carbon neutrality requires efficient, cost-effective carbon capture and utilization (CCU) pathways. Superstructure optimization enables the systematic evaluation of alternative CCU pathways and identification of the most cost-effective and environmentally sustainable configurations. However, existing superstructure optimization approaches are limited by fixed process parameters, reducing model accuracy and flexibility in evaluating novel CCU supply chains. Therefore, this study aims to develop a surrogate-integrated superstructure optimization framework to design CCU supply chains that maximize economic returns and minimize greenhouse gas (GHG) emissions. The proposed model employed a detailed two-level, block-based structure design that captures process-level nonlinearities through artificial neural network-based surrogate models, with a particular focus on methanol synthesis. Environmental and economic trade-offs were examined across 10–90% GHG reduction targets, showing that the surrogate-based approach allows more flexible and realistic optimization than traditional fixed-parameter models. The findings indicate that surrogate-based optimization sustained or improved profitability while enabling more sustainable carbon and energy pathways. This study offers valuable insights to guide the strategic deployment of CCU technologies in global net-zero transition efforts.

키워드